[test-results] [Bamboo] Asterisk Testing > Asterisk Trunk > #1403 has FAILED (80 tests failed, 15 failures were new). Change made by Matthew Jordan.

Bamboo bamboo at asterisk.org
Sat Jun 22 16:34:37 CDT 2013


-----------------------------------------------------------------------
Asterisk Testing > Asterisk Trunk > #1403 failed.
-----------------------------------------------------------------------
Code has been updated by Matthew Jordan.
1/2 jobs failed, with 80 failing tests, 15 failures were new.

http://bamboo.asterisk.org/browse/TESTING-ASTERISKTRUNK-1403/


--------------
Failing Jobs
--------------
  - Asterisk CentOS 6 32-Bit (CentOS 6): 80 of 566 tests failed.



--------------
Code Changes
--------------
Matthew Jordan (392564):

>Fix a deadlock and possible crash in res_fax
>
>This patch fixes two bugs.
>(1) It unlocks the channel in the framehook handlers before attempting to grab
>    the peer from the bridge. The locking order for the bridging framework is
>    bridge first, then channel - having the channel locked while attempting to
>    obtain the bridge lock causes a locking inversion and a deadlock. This
>    patch bumps the channel ref count prior to releasing the lock in the
>    framehook to avoid lifetime issues.
>
>    Note that this does expose a subtle problem in framehooks; that is,
>    something could modify the framehook list while we are executing, causing
>    issues in the framehook list traversal that the callback executes in.
>    Fixing this is a much larger problem that is beyond the scope of this
>    patch - (a) we already unlock the channel in this particular framehook
>    and we haven't run into a problem yet (as modifying the framehook list
>    when a channel is about to perform a fax gateway would be a very odd
>    operation) and (b) migrating to an ao2 container of framehooks would be
>    more invasive at this point. See the referenced ASTERISK issue for more
>    information.
>(2) Directly packing channel variables into a JSON object turned out to be
>    unsafe. A condition existed where the strings in the JSON blob were no
>    longer safe to be accessed if the channel object itself was disposed of.
>
>(issue ASTERISK-21951)
>



--------------
Tests
--------------
New Test Failures (15)
   - AsteriskTestSuite: S/fax/voxfax g711
   - AsteriskTestSuite: S/udptl v6
   - AsteriskTestSuite: S/fax/gateway t38 g711
   - AsteriskTestSuite: S/fax/gateway g711 t38
   - AsteriskTestSuite: S/fax/gateway native t38 ced
   - AsteriskTestSuite: S/udptl
   - AsteriskTestSuite: S/fax/gateway timeout5
   - AsteriskTestSuite: S/fax/gateway timeout1
   - AsteriskTestSuite: S/fax/gateway mix4
   - AsteriskTestSuite: S/fax/gateway mix3
   - AsteriskTestSuite: S/fax/gateway mix2
   - AsteriskTestSuite: S/fax/gateway mix1
   - AsteriskTestSuite: S/fax/gateway no t38
   - AsteriskTestSuite: S/fax/gateway native t38
   - AsteriskTestSuite: S/channels/ s i p/ s d p offer answer
Existing Test Failures (65)
   - AsteriskTestSuite: S/apps/queues/position priority maxlen
   - AsteriskTestSuite: S/apps/voicemail/leave voicemail priority
   - AsteriskTestSuite: S/apps/dial/dial congestion
   - AsteriskTestSuite: S/apps/dial/dial parallel/dial parallel single unavail
   - AsteriskTestSuite: S/bridge/blindxfer setup
   - AsteriskTestSuite: S/channels/ s i p/sip blind transfer/callee with reinvite
   - AsteriskTestSuite: S/channels/ s i p/sip hold
   - AsteriskTestSuite: S/bridge/transfer failure
   - AsteriskTestSuite: S/apps/dial/dial parallel/dial parallel single congestion
   - AsteriskTestSuite: S/apps/voicemail/leave voicemail forwarding

--
This message is automatically generated by Atlassian Bamboo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/test-results/attachments/20130622/33e4904a/attachment-0001.htm>


More information about the Test-results mailing list